Patchwork

Getting started
- Installation
Above & Below
Grouping & Nesting
Layout
Annotation

# Packages you need
library(tidyverse)

# Packages you may not have and can install
library(ggrepel)

Getting started

Installation

Patchwork is not yet available on CRAN. You’ll need to install it as follows.

# install.packages("devtools")
# devtools::install_github("thomasp85/patchwork")

library(patchwork)
load('data/starwars_df.RData')  # load starwars data frames

bar = people_df %>% 
  mutate(homeworld = fct_lump(homeworld, n = 4)) %>% 
  drop_na(homeworld) %>% 
  ggplot() + 
  geom_bar(aes(x = homeworld, fill = homeworld), show.legend = F)

smooth = people_df %>% 
  filter(!is.na(gender) & !grepl(name, pattern = 'Jabba')) %>% 
  ggplot(aes(x=mass, y=height)) + 
  geom_point() + 
  geom_smooth()

scatter = vehicles_df %>% 
  ggplot(aes(x=length, y=cost_in_credits)) + 
  geom_point(size = 10, color = '#ff5500') +
  geom_text_repel(aes(label = name), size=6, color = 'gray50')  # requires ggrepel

box = people_df %>% 
  mutate(species = fct_lump(species, n = 2)) %>% 
  drop_na(species) %>% 
  ggplot(aes(x= species, y = mass)) +
  geom_boxplot()

Above & Below

Basics. A good way to start is that you create the two separate plots and plotting side by side is the first step. Either of these will work.

bar + smooth  # note this could be the actual underlying code of the plots

bar | box

If we want the second plot below, you can just use a divisor.

bar / smooth

Grouping & Nesting

If we have multiple plots, we may want to group them a certain way.

(bar + smooth) / box

This is what it would look like if we hadn’t nested. Operations take similar mathematical precedence.

bar + smooth / box

We can take this further to use nesting to further enhance the plot.

bar / {
  box + {
    scatter / 
      smooth
    }
  }

Layout

Layouts provide more customization over the look.

box + smooth + bar + scatter +
  plot_layout(ncol = 2)

We have additional options.

box + 
  smooth + 
  bar + 
  scatter +
  plot_layout(
    ncol = 2,            # 2 columns
    byrow = T            # enter plots by row
  )

We can also use it with nesting.

bar + {
  box + 
    scatter + 
    plot_layout(ncol = 1)
  } + 
  smooth

Sizing

We can alter relative sizes of plots.

box + 
  smooth + 
  bar + 
  scatter +
  plot_layout(
    ncol = 2,            # 2 columns
    byrow = T,           # enter plots by row
    widths = c(1,2),     # relative column widths
    heights = c(1,1.5)   # relative row heights
  )

Spacing

Add some space between plots. This is all it does at present.

box + 
  plot_spacer() +
  smooth

Applying modifications to all plots

You may have some layouts or themes you’d like to apply to each plot. First we’ll start with the usual approach.

box + smooth + bar + theme_bw()

We could do the following, but it’s tedious.

box + theme_bw() + smooth + theme_bw() + bar + theme_bw()

Instead, we can just nest/group the plots and use * to apply the theme to all. This is much easier.

(box + smooth + bar) * theme_bw()

Using & will apply through all levels of nesting.

box / {
  smooth + 
    bar
  } * theme_bw()



box / {
  smooth + 
    bar
  } & theme_bw()

Annotation

Annotating applies only on the ‘top-level’ plot.

# plot annotation must be applied to the whole set of plots. I nested everything
# so that I could also use the `&` operator for a common theme.
{box / {
  smooth + 
    bar
  }   & 
  theme_bw()
  } + 
  plot_annotation(title = 'Star Wars Plots', 
                  subtitle = 'Shenanigans', 
                  theme = theme(title = element_text(size= 24)))

LS0tCnRpdGxlOiAiUGF0Y2h3b3JrIgpvdXRwdXQ6IAogIGh0bWxfbm90ZWJvb2s6IAogICAgaGlnaGxpZ2h0OiBweWdtZW50cwogICAgdGhlbWU6IHNhbmRzdG9uZQogICAgdG9jOiB5ZXMKZWRpdG9yX29wdGlvbnM6IAogIGNodW5rX291dHB1dF90eXBlOiBpbmxpbmUKLS0tCgpgYGB7ciBzZXR1cCwgaW5jbHVkZT1GQUxTRX0KIyBvcHRpb25zIGZvciB0aGUga25pdHRlZCBodG1sIGRvY3VtZW50CmtuaXRyOjpvcHRzX2NodW5rJHNldCgKICBlY2hvPVQsIAogIG1lc3NhZ2UgPSBGLCAKICB3YXJuaW5nID0gRiwKICBjb21tZW50ID0gTkEsCiAgUi5vcHRpb25zPWxpc3Qod2lkdGg9MTIwKSwgCiAgY2FjaGUucmVidWlsZD1GLCAKICBjYWNoZT1ULAogIGZpZy5hbGlnbj0nY2VudGVyJywgCiAgZmlnLmFzcCA9IC43LAogIGRldiA9ICdzdmcnLCAKICBkZXYuYXJncz1saXN0KGJnID0gJ3RyYW5zcGFyZW50JykKKQpgYGAKCmBgYHtyIHBhY2thZ2VzfQojIFBhY2thZ2VzIHlvdSBuZWVkCmxpYnJhcnkodGlkeXZlcnNlKTsgbGlicmFyeShwYXRjaHdvcmspCgojIFBhY2thZ2VzIHlvdSBtYXkgbm90IGhhdmUgYW5kIGNhbiBpbnN0YWxsCmxpYnJhcnkoZ2dyZXBlbCkKYGBgCgoKIyMgR2V0dGluZyBzdGFydGVkCgojIyMgSW5zdGFsbGF0aW9uCgpgUGF0Y2h3b3JrYCBpcyBub3QgeWV0IGF2YWlsYWJsZSBvbiBDUkFOLiBZb3UnbGwgbmVlZCB0byBpbnN0YWxsIGl0IGFzIGZvbGxvd3MuCgpgYGB7ciBpbnN0YWxsfQojIGluc3RhbGwucGFja2FnZXMoImRldnRvb2xzIikKIyBkZXZ0b29sczo6aW5zdGFsbF9naXRodWIoInRob21hc3A4NS9wYXRjaHdvcmsiKQpgYGAKCgoKCmBgYHtyIGJhc2ljX3Bsb3RzfQpsb2FkKCdkYXRhL3N0YXJ3YXJzX2RmLlJEYXRhJykgICMgbG9hZCBzdGFyd2FycyBkYXRhIGZyYW1lcwoKYmFyID0gcGVvcGxlX2RmICU+JSAKICBtdXRhdGUoaG9tZXdvcmxkID0gZmN0X2x1bXAoaG9tZXdvcmxkLCBuID0gNCkpICU+JSAKICBkcm9wX25hKGhvbWV3b3JsZCkgJT4lIAogIGdncGxvdCgpICsgCiAgZ2VvbV9iYXIoYWVzKHggPSBob21ld29ybGQsIGZpbGwgPSBob21ld29ybGQpLCBzaG93LmxlZ2VuZCA9IEYpCgpzbW9vdGggPSBwZW9wbGVfZGYgJT4lIAogIGZpbHRlcighaXMubmEoZ2VuZGVyKSAmICFncmVwbChuYW1lLCBwYXR0ZXJuID0gJ0phYmJhJykpICU+JSAKICBnZ3Bsb3QoYWVzKHg9bWFzcywgeT1oZWlnaHQpKSArIAogIGdlb21fcG9pbnQoKSArIAogIGdlb21fc21vb3RoKCkKCnNjYXR0ZXIgPSB2ZWhpY2xlc19kZiAlPiUgCiAgZ2dwbG90KGFlcyh4PWxlbmd0aCwgeT1jb3N0X2luX2NyZWRpdHMpKSArIAogIGdlb21fcG9pbnQoc2l6ZSA9IDEwLCBjb2xvciA9ICcjZmY1NTAwJykgKwogIGdlb21fdGV4dF9yZXBlbChhZXMobGFiZWwgPSBuYW1lKSwgc2l6ZT02LCBjb2xvciA9ICdncmF5NTAnKSAgIyByZXF1aXJlcyBnZ3JlcGVsCgpib3ggPSBwZW9wbGVfZGYgJT4lIAogIG11dGF0ZShzcGVjaWVzID0gZmN0X2x1bXAoc3BlY2llcywgbiA9IDIpKSAlPiUgCiAgZHJvcF9uYShzcGVjaWVzKSAlPiUgCiAgZ2dwbG90KGFlcyh4PSBzcGVjaWVzLCB5ID0gbWFzcykpICsKICBnZW9tX2JveHBsb3QoKQpgYGAKCiMjIEFib3ZlICYgQmVsb3cKCkJhc2ljcy4gIEEgZ29vZCB3YXkgdG8gc3RhcnQgaXMgdGhhdCB5b3UgY3JlYXRlIHRoZSB0d28gc2VwYXJhdGUgcGxvdHMgYW5kIHBsb3R0aW5nIHNpZGUgYnkgc2lkZSBpcyB0aGUgZmlyc3Qgc3RlcC4gIEVpdGhlciBvZiB0aGVzZSB3aWxsIHdvcmsuCgoKYGBge3Igc2lkZV9ieV9zaWRlfQpiYXIgKyBzbW9vdGggICMgbm90ZSB0aGlzIGNvdWxkIGJlIHRoZSBhY3R1YWwgdW5kZXJseWluZyBjb2RlIG9mIHRoZSBwbG90cwpgYGAKCmBgYHtyIHNpZGVfYnlfc2lkZTJ9CmJhciB8IGJveApgYGAKCgoKSWYgd2Ugd2FudCB0aGUgc2Vjb25kIHBsb3QgYmVsb3csIHlvdSBjYW4ganVzdCB1c2UgYSBkaXZpc29yLgoKYGBge3IgdG9wX2JvdHRvbX0KYmFyIC8gc21vb3RoCmBgYAoKCiMjIEdyb3VwaW5nICYgTmVzdGluZwoKSWYgd2UgaGF2ZSBtdWx0aXBsZSBwbG90cywgd2UgbWF5IHdhbnQgdG8gZ3JvdXAgdGhlbSBhIGNlcnRhaW4gd2F5LgoKYGBge3IgZ3JvdXBpbmd9CihiYXIgKyBzbW9vdGgpIC8gYm94CmBgYAoKVGhpcyBpcyB3aGF0IGl0IHdvdWxkIGxvb2sgbGlrZSBpZiB3ZSBoYWRuJ3QgbmVzdGVkLiBPcGVyYXRpb25zIHRha2Ugc2ltaWxhcgptYXRoZW1hdGljYWwgcHJlY2VkZW5jZS4KCmBgYHtyIGdyb3VwaW5nMn0KYmFyICsgc21vb3RoIC8gYm94CmBgYAoKV2UgY2FuIHRha2UgdGhpcyBmdXJ0aGVyIHRvIHVzZSAqbmVzdGluZyogdG8gZnVydGhlciBlbmhhbmNlIHRoZSBwbG90LgoKYGBge3IgbmVzdGluZ30KYmFyIC8gewogIGJveCArIHsKICAgIHNjYXR0ZXIgLyAKICAgICAgc21vb3RoCiAgICB9CiAgfQpgYGAKCiMjIExheW91dAoKTGF5b3V0cyBwcm92aWRlIG1vcmUgY3VzdG9taXphdGlvbiBvdmVyIHRoZSBsb29rLgoKYGBge3IgbGF5b3V0MX0KYm94ICsgc21vb3RoICsgYmFyICsgc2NhdHRlciArCiAgcGxvdF9sYXlvdXQobmNvbCA9IDIpCmBgYAoKV2UgaGF2ZSBhZGRpdGlvbmFsIG9wdGlvbnMuCgpgYGB7ciBsYXlvdXQyfQpib3ggKyAKICBzbW9vdGggKyAKICBiYXIgKyAKICBzY2F0dGVyICsKICBwbG90X2xheW91dCgKICAgIG5jb2wgPSAyLCAgICAgICAgICAgICMgMiBjb2x1bW5zCiAgICBieXJvdyA9IFQgICAgICAgICAgICAjIGVudGVyIHBsb3RzIGJ5IHJvdwogICkKYGBgCgoKCldlIGNhbiBhbHNvIHVzZSBpdCB3aXRoIG5lc3RpbmcuCgpgYGB7ciBsYXlvdXQzfQpiYXIgKyB7CiAgYm94ICsgCiAgICBzY2F0dGVyICsgCiAgICBwbG90X2xheW91dChuY29sID0gMSkKICB9ICsgCiAgc21vb3RoCmBgYAoKIyMjIFNpemluZwoKV2UgY2FuIGFsdGVyIHJlbGF0aXZlIHNpemVzIG9mIHBsb3RzLgoKYGBge3Igc2l6aW5nfQpib3ggKyAKICBzbW9vdGggKyAKICBiYXIgKyAKICBzY2F0dGVyICsKICBwbG90X2xheW91dCgKICAgIG5jb2wgPSAyLCAgICAgICAgICAgICMgMiBjb2x1bW5zCiAgICBieXJvdyA9IFQsICAgICAgICAgICAjIGVudGVyIHBsb3RzIGJ5IHJvdwogICAgd2lkdGhzID0gYygxLDIpLCAgICAgIyByZWxhdGl2ZSBjb2x1bW4gd2lkdGhzCiAgICBoZWlnaHRzID0gYygxLDEuNSkgICAjIHJlbGF0aXZlIHJvdyBoZWlnaHRzCiAgKQpgYGAKCiMjIyBTcGFjaW5nCgpBZGQgc29tZSBzcGFjZSBiZXR3ZWVuIHBsb3RzLiBUaGlzIGlzIGFsbCBpdCBkb2VzIGF0IHByZXNlbnQuCgpgYGB7ciBzcGFjaW5nfQpib3ggKyAKICBwbG90X3NwYWNlcigpICsKICBzbW9vdGggCmBgYAojIyMgQXBwbHlpbmcgbW9kaWZpY2F0aW9ucyB0byBhbGwgcGxvdHMKCllvdSBtYXkgaGF2ZSBzb21lIGxheW91dHMgb3IgdGhlbWVzIHlvdSdkIGxpa2UgdG8gYXBwbHkgdG8gZWFjaCBwbG90LiAgRmlyc3Qgd2UnbGwgc3RhcnQgd2l0aCB0aGUgdXN1YWwgYXBwcm9hY2guCgoKYGBge3IgYXBwbHlfdGhlbWUxfQpib3ggKyBzbW9vdGggKyBiYXIgKyB0aGVtZV9idygpCmBgYAoKV2UgY291bGQgZG8gdGhlIGZvbGxvd2luZywgYnV0IGl0J3MgdGVkaW91cy4KCmBgYHtyIGFwcGx5X3RoZW1lMn0KYm94ICsgdGhlbWVfYncoKSArIHNtb290aCArIHRoZW1lX2J3KCkgKyBiYXIgKyB0aGVtZV9idygpCmBgYAoKSW5zdGVhZCwgd2UgY2FuIGp1c3QgbmVzdC9ncm91cCB0aGUgcGxvdHMgYW5kIHVzZSBgKmAgdG8gYXBwbHkgdGhlIHRoZW1lIHRvIGFsbC4gVGhpcyBpcyBtdWNoIGVhc2llci4gIAoKYGBge3IgYXBwbHlfdGhlbWUzfQooYm94ICsgc21vb3RoICsgYmFyKSAqIHRoZW1lX2J3KCkKYGBgCgpVc2luZyBgJmAgd2lsbCBhcHBseSB0aHJvdWdoIGFsbCBsZXZlbHMgb2YgbmVzdGluZy4KCmBgYHtyIGFwcGx5X3RoZW1lNH0KYm94IC8gewogIHNtb290aCArIAogICAgYmFyCiAgfSAqIHRoZW1lX2J3KCkKCgpib3ggLyB7CiAgc21vb3RoICsgCiAgICBiYXIKICB9ICYgdGhlbWVfYncoKQpgYGAKCiMjIEFubm90YXRpb24KCkFubm90YXRpbmcgYXBwbGllcyBvbmx5IG9uIHRoZSAndG9wLWxldmVsJyBwbG90LgoKYGBge3IgYW5ub3RhdGlvbn0KIyBwbG90IGFubm90YXRpb24gbXVzdCBiZSBhcHBsaWVkIHRvIHRoZSB3aG9sZSBzZXQgb2YgcGxvdHMuIEkgbmVzdGVkIGV2ZXJ5dGhpbmcKIyBzbyB0aGF0IEkgY291bGQgYWxzbyB1c2UgdGhlIGAmYCBvcGVyYXRvciBmb3IgYSBjb21tb24gdGhlbWUuCntib3ggLyB7CiAgc21vb3RoICsgCiAgICBiYXIKICB9ICAgJiAKICB0aGVtZV9idygpCiAgfSArIAogIHBsb3RfYW5ub3RhdGlvbih0aXRsZSA9ICdTdGFyIFdhcnMgUGxvdHMnLCAKICAgICAgICAgICAgICAgICAgc3VidGl0bGUgPSAnU2hlbmFuaWdhbnMnLCAKICAgICAgICAgICAgICAgICAgdGhlbWUgPSB0aGVtZSh0aXRsZSA9IGVsZW1lbnRfdGV4dChzaXplPSAyNCkpKSAKYGBgCgo=